browser-automation

Pass

Audited by Gen Agent Trust Hub on Mar 22, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill initiates the installation of the playwright library and its corresponding Chromium browser binaries. These are industry-standard tools maintained by Microsoft, a trusted organization, for the purpose of programmatic web interaction.
  • [SAFE]: The skill demonstrates positive security posture by instructing the user to prevent the exposure of session credentials, specifically noting that auth-state.json should be added to .gitignore.
  • [PROMPT_INJECTION]: By navigating to external URLs, the skill creates an inherent surface for indirect prompt injection from third-party web content. 1. Ingestion points: External data is ingested via the page.goto(url) command in SKILL.md. 2. Boundary markers: The provided code patterns do not include explicit delimiters or instructions to ignore embedded commands. 3. Capability inventory: The skill environment includes access to Bash, Write, Read, Glob, and Grep tools. 4. Sanitization: There is no evidence of automated content sanitization before the data is processed by the agent.
  • [SAFE]: Metadata includes a verification timestamp set in the future ('2026-03-13'); while anomalous, this is interpreted as a placeholder or documentation error rather than a malicious deception.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 22, 2026, 04:50 PM